MARA

AI Memo

MARA

As of 2026-09-03/Confidence medium

Open ticker Back to Memos

Data freshness

AI Memo

As of 2026-09-03 • Model: RankAlpha Sentiment Codex

Decision Panel

Action, confidence, horizon, and decision blockers at a glance.

Action

—

Confidence

medium

Horizon

—

Key catalyst

Q2 earnings reset the near-term thesis • 2026-08-06

Base target: $10 / p=40% prob

Biggest risk

Revenue contracted, adjusted EBITDA turned negative, bitcoin mark-to-market losses were material, energy cost per BTC increased, and the business remains exposed to bitcoin prices, treasury sales, dilution, and execution risk. Recent officer and director open-market sales add a sentiment overhang but are not evidence of operating performance. [#SEC-FORM4-2026-08-19]

Invalidation

Break below $7 or thesis failure.

What would change my mind?

Hashrate reached 70.3 EH/s, up 22% year over year, while cost per petahash per day improved 4%. Successful conversion of power assets into contracted AI/HPC capacity could create material upside, but this remains unproven.

Research note

Recommendation

—

Confidence

medium

As of

2026-09-03

Base target

$10

p=40%

Bull target

$15

p=20%

Bear target

p=40%

Key bullets

Fast scan of bull, bear, and watch items

BullKey bullets

Bitcoin and treasury volatility remain dominant drivers
AI/HPC infrastructure conversion remains option value
Hashrate reached 70.3 EH/s, up 22% year over year, while cost per petahash per day improved 4%.

BearKey bullets

AI/HPC plans lack disclosed tenants, utilization, contracted economics, and clear capex timing.
BTC loaning, collateral pledges, and sales create liquidity, counterparty, and dilution-related risks.
Bitcoin-price volatility and digital-asset mark-to-market losses can materially affect earnings and equity value.

WatchKey bullets

Q2 earnings reset the near-term thesis

Summary

High-level snapshot

Recommendation

Model confidence

medium

Timeframe

Asset

Equity

Trigger

Base case: $10 (p=40%)

Target spread vs base: +58% / -32%

Scenario probabilities

Bull Target: $15 (p=20%)Base Target: $10 (p=40%)Bear Target: $7 (p=40%)

Commentary

News tone is mixed-to-negative: the Q2 company filing confirms a substantial earnings deterioration despite hashrate growth, while stored secondary coverage reports a sharp recent share-price reset. No reliable analyst target-change, estimate-revision, or consensus-surprise dataset is available. Social coverage is unavailable, so confidence remains limited despite strong filing evidence. [#SEC-8K-2026-08-06]

As of 2026-09-03 • Updated bi-weekly / post-earningsSource: Internal modelMethodology

Analyst Summary

Thesis highlights from the latest analysis

Bull Case

Bear Case

Drivers

Scannable catalysts and supporting factors

Bitcoin and treasury volatility remain dominant drivers
AI/HPC infrastructure conversion remains option value

Scenarios

Bull Target (Scenario probability)

$15 / p=20% prob

Base Target (Scenario probability)

$10 / p=40% prob

Bear Target (Scenario probability)

$7 / p=40% prob

Analysis 2026-09-03 · RankAlpha Sentiment Codex

Analysis reference price unavailable. Targets are shown without scenario returns.

Sentiment & Exposure

News Sentiment (30d)

40.00

Social Sentiment (7d)

0.00

Events

Other notable items detected

EVENT2026-08-06
Q2 earnings reset the near-term thesis
Q2 2026 revenue was $174.9 million, down 27% year over year; net loss was $611.3 million and adjusted EBITDA was negative $360.9 million. Hashrate growth and 2,422 BTC mined provided operating positives, but consensus-surprise data is unavailable. [#SEC-8K-2026-08-06]
Prob 100.00% • Move -12.00% • Priced-in 75.00%

Short-term Catalysts

Next few weeks

Bitcoin and treasury volatility remain dominant drivers
2026-12-02
MARA held 35,577 BTC at June 30, 2026, including 9,270 BTC loaned or pledged as collateral, produced 2,422 BTC, and sold 2,213 BTC in Q2. Purchased energy cost per BTC rose to $38,690 from $33,735, leaving earnings and liquidity highly sensitive to bitcoin prices and operating costs. [#SEC-8K-2026-08-06]
Prob 75% • Move -10% • Priced-in 65%

Long-term Catalysts

Months to a year

AI/HPC infrastructure conversion remains option value
2026-09-30
The Starwood partnership was described as supporting approximately 1 GW of near-term IT capacity with a pathway above 2.5 GW, but disclosed tenants, utilization, capex cadence, and contracted economics remain limited. A rerating therefore requires harder operating proof. [#PR-2026-02-26-Starwood]
Prob 30% • Move 25% • Priced-in 25%

Risks

AI/HPC plans lack disclosed tenants, utilization, contracted economics, and clear capex timing.
BTC loaning, collateral pledges, and sales create liquidity, counterparty, and dilution-related risks.
Bitcoin-price volatility and digital-asset mark-to-market losses can materially affect earnings and equity value.
Purchased energy cost per bitcoin increased to $38,690, creating margin pressure.
Q2 revenue declined 27% year over year and adjusted EBITDA was negative $360.9 million.
Recent officer and director stock sales may weigh on sentiment. [#SEC-FORM4-2026-08-19]

Peers

Symbol	Summary
RIOT	Close operating peer combining bitcoin mining with AI/HPC evaluation of power assets; useful for assessing whether power and data-center optionality can support a rerating.
CLSK	Cleaner pure-mining benchmark for hashrate execution and treasury discipline, but less relevant to AI optionality.
IREN	Stronger hybrid benchmark because it has disclosed secured power and a Microsoft AI-cloud agreement, evidence MARA still lacks for its Starwood narrative.
BTDR	Hybrid mining and AI-cloud peer reporting AI cloud ARR and tenant negotiations, highlighting MARA's need for utilization and revenue proof.

Sources

Internal and external references used in this analysis

Model source: RankAlpha Sentiment Codex
Analysis date: 2026-09-03

Loading AI feed…

Symbol

Summary

RIOT

Close operating peer combining bitcoin mining with AI/HPC evaluation of power assets; useful for assessing whether power and data-center optionality can support a rerating.

CLSK

Cleaner pure-mining benchmark for hashrate execution and treasury discipline, but less relevant to AI optionality.

IREN

Stronger hybrid benchmark because it has disclosed secured power and a Microsoft AI-cloud agreement, evidence MARA still lacks for its Starwood narrative.

BTDR

Hybrid mining and AI-cloud peer reporting AI cloud ARR and tenant negotiations, highlighting MARA's need for utilization and revenue proof.